Wireless communication systems that provide communication services to wireless communication devices have evolved during the last decade to meet the ever increasing demand for capabilities related to high speed data transfer and to provide these capabilities at any given time and at any geographical location. In order to provide the services at any time and place, the wireless networks typically comprise one or more network nodes, such as radio base stations, that maintain one or more radio cells and the wireless communication devices communicate via these network nodes.
However, in some scenarios wireless networks suffer from the problems of having a limited range and unreliable communications such that data is sometimes received corrupted at the receiver end. As a means to mitigate such range limitations, so-called mesh networks have been developed in which intermediary relay nodes relay data from a source node to a destination node and hence extends the range of the nodes in the mesh network. In other words, in a mesh network wireless communication devices may operate as relay nodes and communicate directly with each other without involvement from a radio base station.
Connectivity of a mesh network describes the ratio with which network nodes can reach another node in the mesh network. In the case of a fully connected mesh network all nodes are able to reach all other nodes. Often, for example in machine type communication, MTC, involving sensors or meters, communication in the mesh network is performed towards a single recipient which is a server receiving sensor or metering data. As a consequence, the only relevant connection on an applications level for a node in such a mesh network is with the server receiving its data. In this case the network is fully connected if all nodes can communicate with the server. For such a mesh network, a proactive routing technique, in which paths are maintained by periodic signaling, is highly suitable since only one route needs to be maintained for all nodes. Hence, if a node is able to find another node connected to the server it may itself connect to the server through that node. Alternatively, for more arbitrarily chosen destination nodes, a reactive network may be more suitable. In such a case a path is set up when a communication need arises, and the link dies sometime after the communication has ceased.
Path setup in a mesh network is well known in the art. For example, in the Institute of Electrical and Electronics Engineers, IEEE, standard 802.11™-2012, a source node broadcasts a Path Request, PREQ, in order to establish a route from itself to a destination node. The PREQ is relayed by all nodes receiving it correctly until it reaches the destination node. Upon receiving the PREQ, the destination node answers with a unicast Path Reply, PREP, to its closest relay node. The PREP is then forwarded along the relay nodes until it reaches the source node and the path is established. The source node then transmits data to the first node in a list of relay nodes that defines the path. The header of a data message typically contains four address fields: a source address and a destination address representing the end points in the path and also a transmitter address and a receiver address, representing end points of a present link between two nodes in the path.
Irrespective of whether the mesh network is reactive or proactive, a path is set up between the source and destination nodes within the mesh network. In order for a message to be forwarded from a source node to destination node, the message only needs to contain information about its final destination. From that information the node is able to look up the node next in line, i.e., which node to relay the message to for it to eventually reach the destination.
Paths in mesh networks are often set up in order to minimize a path metric, e.g. Airtime Link Metric, ATLM, in order to establish an optimal path. Such a metric often takes into account the use of sparse radio resources but not the processing time required for each node to forward a message or, in a worst case, to ask for a retransmission of a message. Neither does such a metric consider the dynamics of wireless communications which are very volatile such that only a small change in the environment may cause a large difference in channel characteristics. What at one time is the best path according to some metric may not be best at another time. For example, a relay node that is unable to receive a message at one point in time may be able to receive a message at another point in time due to the fact that mesh networks may be dynamic in the sense that individual nodes may move within a mesh network. A node that for some reason falls outside its mesh network is detached and hence unable to communicate. In other words, prior art message relay is often burdened with unnecessarily high latency.